Computer Simulation Of Vegetation Canopy At Multi-scale And Sensitivity Analysis Of Parameters

The researches on vegetation BRDF (Bi-directional reflectance distribution function) models are the theoretical basis to support multi-angle remote sensing techniques of vegetations, which promote the process of quantitative remote sensing and apply to many fields, such as military, and space remote sensing, meteorological and ecological environment fields. Therefore, vegetation BRDF models (such as, Radiation Translation(RT) model, Geometry Optical(GOMS), model GORT model and Computer Simulation model have become the hot field in remote sensing science to calculate vegetation canopy bi-directional reflectance .In this study, Radiosity method is used to simulate the canopy BRF (bi-directional reflectance factor) of vegetation, which is one of the methods of computer simulation model. The Radiosity method composes of two modules: 1) using L-system to render the 3D realistic structured scenes; 2) based on the rendering scenes, simulating the canopy bi-directional reflectance using Radiosity method. In order to successfully carry out researches on the computer simulation model, some experiments have been done, and accumulated lots of experimental data, which are the conditional data for model study, and also are the validation authority for simulation results.In this paper, taking crops (winter wheat and summer maize) and broadleaf trees as research objects, their canopies BRF have been simulated by Radiosity method. In addition, in order to extend the simulated scale for Radiosity method, the crown structures are simplified. In this paper, the major research contents and innovations points are as following:(1) Based on the measured leaf area index(LAI) and relative cumulated temperature, and used Logistic function, the model is fitted to depict the LAI changes in the entire crop growth stages. Taking LAI as the control factor, we simulate canopy bi-directional reflectance and NDVI for different LAI, and analyzing their change. Computer simulated results(canopy BRF and hyperspectrum data) are compared with field measured data ,Prosail simulated data and MISR BRF data, and they have made good consistency.(2) Based on canopy BRF simulation of winter wheat, studying on the influence of winter wheat stem to canopy bi-directional reflectance regime in different growth stages, and analyzing the effect of stem on the day change of canopy bi-directional reflectance. In addition, the realistic structured scenes for two different ridges are simulated. The study shows that the ridge has some influence on canopy BRF of winter wheat.(3) In order to analyze the influence of underlying surface on crop canopy bi-directional reflectance regime, two mixed scenes of maize and grass with different types of underlying surfaces are simulated: one is soil and sparse grass, and the other is soil and bushy grass. One conclusion is shown that the influence of underlying surface with bushy grass is great, which can accumulate the priori knowledge to reverse the structure parameters.(4) In order to extent the scale of realistic structured scene for computer simulation study, forest is taken as the experiment object. Based on the simulated realistic structured broadleaf forest scene, canopy radiation regime is analyzed. In addition, we also study on the change trend of canopy BRF for different LAI.(5) Because of the complexity of crown 3D structures, and the facets are too many to calculate canopy BRF by radiosity method. So in order to extent the simulated scale and solve the problem of larger scale simulation, an idea is to abstract the crown as an ellipsoid and to simplify the scene structures. When facets in the scenes are given optical properties, the probability of crown gap is considered, and combined with the thoughts of Geometry Optical model to complete canopy BRF simulation at larger scale.(6) Take advantage of computer simulation model, based on vegetation scenes in many different conditions, analyzing the sensitivity of parameters to vegetation canopy bi-directional reflectance regime. The parameters conclude the spatial distribution, sun direction and underlying surface in simulated scenes.